Why design matters for a web framework: a 7-year evolution | Wasp

Skip to main content<br>Wasp is now 100% TypeScriptMeet TS Spec →

Wasp is a full-stack web framework for React and Node.js (imagine Rails for TypeScript, with but with frontend included and wired up out-of-the-box), and we've been building it for a little over five years (as a full-time thing). For almost all of that time, we saw design and "branding" as something we shouldn't spend too much time on.

A framework is mostly code, we thought: you drive it from a CLI and your editor, there is little visual to look at, and these days an agent reads the docs for you half the time anyway. The most loved frameworks kind of prove the point, with Django and Rails famous and famously plain.

Turns out we were wrong, though not in the way you'd expect. What made design start to matter wasn't users or growth - it was us, the people building Wasp and the community around it.

This is the story of how we figured that out, and of the rebrand it led to, which shipped as part of Launch Week #12 - the same week Wasp went fully TypeScript-native.

Five years of not having a design​

Same product, five years apart. Very different vibes.

Wasp has always had a logo, colors and fonts (ok, the fonts were whatever shipped with Docusaurus). So when I say we didn't have a design, I mean we never sat down to think about it as a whole. There was no central story, no backbone holding everything together, just a bunch of pages that happened to be ours.

Here's how it typically went down: with every big release (v0, Alpha, Beta), I'd sit down and convince Martin (my twin brother and co-founder) to let me "rebrand" Wasp. And every single time, I'd do almost the exact same thing: look around for some inspiration (thanks, Supabase), then patch, okay, steal, a few sections together into something that looked decent enough, and ship it.

And early on, that was the right call. When you're still proving the product deserves to exist, obsessing over a design system is a great way to burn weeks you don't really have. Plenty of founders do the opposite. They polish pixels on something nobody wants yet (and we witnessed that many times) and we were determined not to be them.

But design debt compounds, just like code debt does. Over five years, Wasp itself grew up a lot, and it grew much faster than its design did, until at some point it just overgrew it. The framework got serious, but the brand stayed improvised. You can kind of watch it happen across our old landing pages:

stic (before it was called wasp), 2019<br>01 / 04

‹›

And once our design felt broken, there was little motivation to keep it right, so it just kept getting worse. At one point we needed to show a roadmap on the site, so Martin just screenshotted our GitHub project board and pasted it straight in. Low-res, rounded corners not quite matching, with some random background bleeding out around the edges. It stayed up like that for an embarrassingly long time:

Our roadmap section. Yes, that is just a screenshot of our GitHub project. Not even properly cropped.

Or take the "How does it work?" section right below it. The diagram was originally drawn as a transparent SVG, which broke once the site got a dark mode, so someone slapped a white background behind it to "fix" it. Now it looked fine on the dark version, and noticeably worse everywhere else. Add the orange highlight pills next to the copy, styled in a completely different visual language than anything else on the page, and the whole thing kind of looked like three different sites stitched together:

Our old "How it works" section, broken in a few different ways.

At some point we just started to feel bad about how unpolished it looked. We couldn't really tell you what users made of it (probably not much), but we could feel it ourselves: Wasp had grown into a serious product that people use for serious things, and it was about time it started looking like one.

Looking unpolished made us feel unpolished​

There was also another reason for finally taking design seriously, harder to pin down, but looking back it's pretty obvious.

Looking unpolished slowly does something to you. Every time we shipped a great feature next to a website that looked like a slightly abandoned student project, it sent a quiet little signal we never really meant to send: that maybe we don't actually care all that much. And that signal doesn't only go out to users; it also comes back in. A product that looks like nobody sweated the details slowly starts to feel like one to the people building it, even when, internally, the opposite is completely true.

You don't really notice it on any given day, but over five years it compounds. You hesitate a bit before posting the screenshot. You add a caveat before the demo ("ignore how it looks for a second, just look at what it does"). You quietly stop wanting to share Wasp with someone whose opinion you actually care about, because they might judge...